Anthony O'Hagan & Jeremy E. Oakley
Department of Probability and Statistics, University of Sheffield, Sheffield, England
Publication details: Reliability Engineering and System Safety 85, 239-248, 2004.
The workshop on Epistemic Probability at Albuquerque, New Mexico, in August 2002 focused on a set of "challenge problems" concerning the expression and propagation of uncertainty based on expert knowledge. This paper considers the general question of elicitation of expert knowledge. We argue that the only logical way to represent uncertainty is via probability. We further argue that the challenge problems are unrepresentative of the kind of probabilistic information that can realistically be elicited from experts. Nevertheless, we address some of the challenge problems as originally posed, using a recently developed Bayesian methodology that acknowledges and quantifies the imprecision in experts' statements.